1. Field of the Invention
The present invention is directed, in general, to wireless communication systems and, more specifically, to a Single-Carrier Frequency Division Multiple Access (SC-FDMA) communication system.
2. Description of the Art
In particular, the present invention is directed to the transmission and multiplexing of Reference Signals (RSs) in SC-FDMA communication systems.
Several types of signals should be supported for the proper functionality of a communication system. In addition to data signals, which convey information content of a communication, control signals also need to be transmitted from User Equipments (UEs) to their serving Base Station (BS or Node B) in the UpLink (UL) of the communication system and from the serving Node B to the UEs in the DownLink (DL) of the communication system in order to enable the proper transmission of data signals. For example, control signals include positive or negative acknowledgement signals (ACK or NACK, respectively), transmitted by a UE in response to (correct or incorrect, respectively) data packet reception and Channel Quality Indication (CQI) signals conveying information about the DL channel conditions experienced by the UE. Furthermore, RSs (also known as pilots) are typically transmitted by each UE having UL data or control transmission. These RSs provide coherent demodulation for the transmitted data and will be referred to as DeModulation (DM) RSs.
The present invention considers the UL communication and assumes that the transmission of signals carrying the data content information from UEs is through a Physical Uplink Shared CHannel (PUSCH) while, in the absence of data information, the transmission of control signals from the UEs is through the Physical Uplink Control CHannel (PUCCH).
A UE, also commonly referred to as a terminal or a mobile station, may be fixed or mobile and may be a wireless device, a cellular phone, a personal computer device, a wireless modem card, etc. Additionally, a Node B is generally a fixed station and may also be called a Base Transceiver System (BTS), an access point, etc.
The UEs are assumed to transmit data or control signals over a Transmission Time Interval (TTI), which in an exemplary embodiment of the present invention corresponds to a sub-frame.
FIG. 1 illustrates a block diagram of a sub-frame structure 110 assumed in an exemplary embodiment of the present invention for PUSCH transmission. The sub-frame includes two slots. A first slot 120 further includes seven symbols used for the transmission of data and/or control signals. Each symbol 130 further includes a Cyclic Prefix (CP) for mitigating interference caused by channel propagation effects. The signal transmission in one slot may be in the same part or it may be at a different part of the operating bandwidth than the signal transmission in the other slot. In addition to symbols carrying data or control information, some symbols may be used for the RS transmission, i.e., DM RS 140, to provide channel estimation and enable coherent demodulation of the received signal. It is also possible for the TTI to include only one slot or more than one sub-frames.
The transmission BandWidth (BW) is assumed to include frequency resource units, which will be referred to herein as Resource Blocks (RBs). An exemplary embodiment of the present invention assumes that each RB includes 12 sub-carriers, and that UEs are allocated a multiple N of consecutive RBs 150 for PUSCH transmission and 1 RB for PUCCH transmission. Nevertheless, it should be noted that the above values are only illustrative and should restrict the described embodiments of the invention.
In order for the Node B to determine the RBs where to schedule the PUSCH transmission by a UE and the Modulation and Coding Scheme (MCS) used for the data, a CQI estimate is needed over the PUSCH transmission BW, which is smaller than or equal to the operating BW. Typically, this CQI estimate is obtained through the transmission by the UE of another RS sounding the scheduling bandwidth (Sounding RS or SRS). This SRS is transmitted in a symbol of an UL sub-frame replacing the data, it is used to provide a Signal-to-Interference and Noise Ratio (SINR) estimate over the RBs comprising its transmission BW, and it can be further used for UL Transmission Power Control (TPC) and UL synchronization.
FIG. 2 illustrates an exemplary embodiment for SRS transmission occurring in one sub-frame symbol every 2 sub-frames for a respective 4.3% SRS overhead. UE1 210 and UE2 220 multiplex their PUSCH transmissions in different parts of the operating BW during a first sub-frame 240, while UE2 220 and UE3 230 multiplex their PUSCH transmissions in different parts of the operating BW during a second sub-frame 250. In some symbols of the sub-frame, the UEs transmit DM RS in order to enable the Node B receiver to perform coherent demodulation of the data signal transmitted in the remaining symbols of the sub-frame with UE1, UE2, and UE3 transmitting respectively DM RS 260, 270, and 280.
In the exemplary structure illustrated in FIG. 2, the first symbol every second sub-frame is used for SRS transmission 290. The UEs having SRS transmission may or may not have PUSCH transmission in the same sub-frame. The SRS transmission may occupy a different part of the operating BW than the data or DM RS transmission from a UE. Moreover, although in the exemplary embodiment the SRS transmission occurs during the first symbol of a sub-frame, any other symbol, such as the last symbol of a sub-frame, may be used.
An exemplary block diagram for data transmission through SC-FDMA signaling is illustrated in FIG. 3. Referring to FIG. 3, the encoded data 310 is provided to a Discrete Fourier Transform (DFT) unit 320, the sub-carriers 330 corresponding to the assigned transmission BW are selected 340, the Inverse Fast Fourier Transform (IFFT) is performed 350, and finally the Cyclic Prefix (CP) 360 and filtering 370, such as time windowing, are applied to the transmitted signal. Zero padding is assumed to be inserted by the reference UE in sub-carriers used for the signal transmission by another UE and in guard sub-carriers (not shown).
Moreover, for brevity, additional transmitter circuitry such as digital-to-analog converters, analog filters, amplifiers, transmitter antennas, etc., are not illustrated in FIG. 3. Similarly, the encoding process and the modulation process for the data, which are well known in the art, such as turbo coding and Quadrature Phase Shift Keying (QPSK), or Quadrature Amplitude Modulation (QAM) 16, or QAM64 modulation, are also omitted for brevity.
FIG. 4 illustrates an exemplary transmitter structure for the DM RS, which is assumed to be based on the time-domain transmission of Constant Amplitude Zero Auto-Correlation (CAZAC) sequences and will be subsequently described in detail.
In FIG. 4, a CAZAC-based sequence 410 is cyclically shifted 420. The DFT of the resulting sequence is obtained 430, the sub-carriers 440 corresponding to the assigned transmission BW are selected 450, the IFFT is performed 460, and finally, the CP 470 and filtering 480 are applied to the transmitted signal 490. Zero padding is assumed to be inserted by the reference UE in sub-carriers used for the signal transmission by another UE and in guard sub-carriers (not shown).
The exemplary transmitter structure illustrated in FIG. 4 can also be used, possibly with minor modifications (such as the repetition in time of the CAZAC-based sequence to produce a comb spectrum), for the SRS transmission.
As for the data transmission, for brevity, additional transmitter circuitry such as digital-to-analog converters, analog filters, amplifiers, transmitter, etc., are not illustrated in FIG. 4.
An alternative generation method for the transmitted CAZAC-based sequence, serving as DM RS or as SRS, is in the frequency domain. This is illustrated in FIG. 5. For the SRS it is also possible that the selected sub-carriers are not consecutive (comb spectrum) in order to multiplex SRS transmissions from multiple UEs over different BWs. However, this is not material to the present invention.
Referring to FIG. 5, the generation of the transmitted CAZAC-based sequence in the frequency domain follows the same steps as in the time domain with two exceptions. The frequency domain version of the CAZAC-based sequence is used 510 (that is, the DFT of the CAZAC-based sequence is pre-computed and not included in the transmission chain) and the cyclic shift 550 is applied after the IFFT 540. The selection 520 of the sub-carriers 530 corresponding to the assigned transmission bandwidth, and the application of CP 560 and filtering 570 to the transmitted signal 580, as well as other conventional functionalities (not shown), are the same as previously described for FIG. 3.
At the receiver, the inverse (or complementary) transmitter functions are performed. For the DM RS, this is conceptually illustrated in FIG. 6, in which the reverse operations of those in FIG. 4 apply, and in FIG. 7, in which the reverse operations of those in FIG. 5 apply.
Referring to FIG. 6, an antenna receives the Radio-Frequency (RF) analog signal and after being processed by further processing units (such as filters, amplifiers, frequency down-converters, and analog-to-digital converters) the digital received signal 610 passes through a time windowing unit 620 and the CP is removed 630. Subsequently, the receiver unit applies an FFT 640, selects 650 the sub-carriers 655 used by the transmitter, applies an Inverse DFT (IDFT) 660, removes the cyclic shift 670 applied to the transmitted CAZAC-based sequence and, using a replica of the CAZAC-based sequence 680, multiplies (correlates) with the resulting signal 690 to produce an output 695 that can be used for channel estimation or CQI estimation for the UL channel.
Similarly in FIG. 7, the digital received signal 710 passes through a time windowing unit 720 and the CP is removed 730. Subsequently, the cyclic shift of the transmitted CAZAC-based sequence is restored 740, an FFF 750 is applied, the selection 760 of the transmitted sub-carriers 765 is performed and correlation 770 with the CAZAC-based sequence replica 780 is subsequently applied. Finally, the output 790 is obtained which can then be passed to a channel estimation unit, such as a time-frequency interpolator, or a CQI estimation unit for the UL channel.
As for the transmitter, functionalities that are well known in the art, such as channel estimation, demodulation, and decoding are not shown for brevity, as they are not material to the present invention.
As mentioned above, the RS (DM RS or SRS) is assumed to be constructed from CAZAC-based sequences. An example of such sequences is given by the following Equation (1):
                                          c            k                    ⁡                      (            n            )                          =                              exp            ⁡                          [                                                                    j2π                    ⁢                                                                                  ⁢                    k                                    L                                ⁢                                  (                                      n                    +                                          n                      ⁢                                                                        n                          +                          1                                                2                                                                              )                                            ]                                .                                    (        1        )            
In Equation (1), L is a length of the CAZAC sequence, n is an index of an element of the sequence n=(0, 1, 2 . . . , L−1), and k is an index of the sequence itself. For a given length L, there are L−1 distinct sequences, if L is prime. Therefore, an entire family of sequences is defined as k ranges in {1, 2 . . . , L−1}. However, it should be noted that the CAZAC sequences used for the CQI and RS generation need not be generated using the exact above expression as will be further discussed below.
For CAZAC sequences of prime length L, the number of sequences is L−1. As the RBs are assumed to include an even number of sub-carriers, with 1 RB including 12 sub-carriers, the sequences used to transmit the ACK/NACK and RS can be generated, in the frequency or time domain, by either truncating a longer prime length (such as length 13) CAZAC sequence or by extending a shorter prime length (such as length 11) CAZAC sequence by repeating its first element(s) at the end (cyclic extension), although the resulting sequences do not specifically fulfill the definition of a CAZAC sequence. Alternatively, CAZAC-based sequences can be generated through a computer search for sequences satisfying the CAZAC properties.
Different cyclic shifts of the same CAZAC sequence provide orthogonal CAZAC sequences. Therefore, different cyclic shifts of the same CAZAC sequence can be allocated to different UEs and achieve orthogonal multiplexing of the RS transmitted by these UEs in the same RBs. This principle is illustrated in FIG. 8.
Referring to FIG. 8, in order for the multiple CAZAC-based sequences 810, 830, 850, and 870 generated correspondingly from multiple cyclic shifts 820, 840, 860, and 880 of the same root CAZAC-based sequence to be orthogonal, the cyclic shift value Δ 890 should exceed the channel propagation delay spread D, including a time uncertainty error and filter spillover effects. If Ts is the duration of one symbol, the number of cyclic shifts is equal to the mathematical floor of the ratio Ts/D. For 12 cyclic shifts and for symbol duration of about 66 microseconds (14 symbols in a 1 millisecond sub-frame), the time separation of consecutive cyclic shifts is about 5.5 microseconds. Alternatively, to provide better protection against multipath propagation, only 6 cyclic shifts should be used, providing time separation of about 11 microseconds.
The present invention assumes that the PUCCH transmission is entirely based on CAZAC-based sequences and, similarly to FIG. 1 for the PUSCH, corresponding exemplary PUCCH structures for ACK/NACK transmission and CQI transmission are respectively illustrated in FIG. 9 and FIG. 10. Also as for PUSCH VoIP transmissions, the PUCCH transmission in the first slot of a sub-frame is assumed to occur at a different part of the operating bandwidth than the PUCCH transmission in the second slot of the sub-frame.
In the exemplary structure illustrated in FIG. 9, a UE transmits ACK/NACK 910 in four symbols of each slot and RS 920 in three symbols of each slot. Both ACK/NACK and RS transmissions are based on the transmission of a CAZAC-based sequence 930 (modulated in case of ACK/NACK by the respective ACK/NACK bits). The multiplexing of ACK/NACK signals and RS from different UEs is through the use of different cyclic shifts of a CAZAC-based sequence, as previously described, and through the use of orthogonal covers {W1, W2, W3, W4} 940 of length 4, such as Walsh-Hadamard codes, for the ACK/NACK and length 3 {H1, H2, H3} 950, such as DFT codes, for the RS.
In the exemplary setup illustrated in FIG. 10, the CQI bits 1010 are transmitted in five symbols of each slot and the respective RS 1020 is transmitted in two symbols of each slot. The multiplexing of CQI signals and RS from different UEs is again through different cyclic shifts of a CAZAC-based sequence (modulated in case of CQI) 1030. An orthogonal cover {G1, G2} 1040 of length 2, such as Walsh-Hadamard, may also apply for the RS transmission.
The transmitter and receiver structures for the CAZAC-based sequences used in PUCCH transmissions are practically the same with the corresponding ones in FIG. 4 and FIG. 5 (transmitter) and FIG. 6 and FIG. 7 (receiver), respectively, and are not repeated for brevity.
An exemplary embodiment of the present invention considers Voice over Internet Protocol (VoIP), which represents an important application in communication systems. Unlike data packets, VoIP packets typically occupy a predetermined BW of a few RBs due to their small size. For BW fragmentation reasons, due to the single carrier property, and for frequency diversity of VoIP transmissions, the RBs may be towards one edge of the operating BW in the first slot and towards the other edge during the second slot of the sub-frame. Also, unlike data packets, VoIP packets are typically transmitted at predetermined time intervals, such as once every 20 msec.
Given that a large number of VoIP UEs, such as 200-300 at 5 MHz system BW or 400-600 at 10 MHz system BW, may be active for a fully loaded system, the respective SRS transmission overhead is an important consideration. For example, if 600 VoIP UEs for 10 MHz system BW are equally divided over a period of 20 msec and assuming a 50% voice activity factor, 15 VoIP UEs will be transmitting in every sub-frame. In several operating environments the number of cyclic shifts that can be used to orthogonally multiplex CAZAC-based sequences transmitted over common BW is much smaller than 15 and substantial overhead is required to support SRS transmissions from VoIP UEs.
Therefore there is a need to support SRS transmission for VoIP UEs without introducing large respective overhead in the UL of the communication system. Moreover, there is a need for the SRS transmission from VoIP UEs to enable optimum sounding conditions in order for the Node B to perform link adaptation. Link adaptation may be in the form of TPC commands or MCS adjustments transmitted by the serving Node B to a VoIP UE.
Additionally, considering the voice activity, it is desirable to re-assign radio resources in the silent period, i.e., when a VoIP UE does not have a data packet to transmit, to another UE in order to improve BW utilization. A VoIP UE may transmit a Release Request (RR) to its serving Node B to indicate that its buffer is empty. Conversely, when a VoIP UE exits a silent period, it needs to send a Service Request (SR) to indicate that it has packets for transmission. This operation is illustrated in FIG. 11 for the case of RR but the same concept applies in case of SR.
Referring to FIG. 11, a VoIP UE transmits packets 1110 and when its buffer empties, the VoIP UE sends a RR 1120. When the serving Node B receives the RR 1130, it re-assigns the BW resources to the VoIP transmission from another UE 1140.
The nature of the single-state transmission information the VoIP UE sends (RR, SR, etc.) is immaterial to the purposes of the present invention. The focus is instead on the signaling used by the VoIP UE to send this single-state information.
The prior art considers that either a separate channel is used for the transmission of the RR or SR, or a few bits are punctured from the CQI feedback the VoIP UE sends to its serving Node B regarding the DL SINR operating conditions and replaced with the transmission of the RR or SR. The first prior art option introduces additional UL overhead and system complexity as a new channel needs to be defined and supported. The second prior art option typically compromises the accuracy of the CQI reception and of the RR or SR reception, and increases the complexity of the Node B receiver.
There is consequently a need to provide a signaling mechanism for VoIP UEs to transmit state transition information, such as a RR or SR, without introducing additional overhead or compromising the quality of other signals such as the CQI, while introducing only minimal additional complexity to the Node B receiver operation.